Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Tech Talks With Kinsoft

Tech Talks With Kinsoft artwork

Last Week in Tech – Apple's Gemini-Powered Siri, ShinyHunters Hit PeopleSoft, and the AI-Chip Land Grab

Jun 14, 202621 minAudio

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Your Monday catch-up on the week that was in tech, with a security lens. This week: Apple finally reveals its overhauled, AI-powered Siri at WWDC 2026 - running partly on Google's Gemini models. The AI infrastructure land grab heats up as Alphabet rents out its new TPUs in a $5B Blackstone joint venture and Microsoft ships its Maia 200 accelerator. Anthropic's Claude posts eye-watering growth and formalises a $100M partner program. And on the security desk: ShinyHunters exploit an Oracle PeopleSoft flaw to hit universities, Europol dismantles a €336M crypto-laundering pipeline used by ransomware crews, CISA flags three newly exploited vulnerabilities, and a new gang called The Gentlemen climbs the ransomware rankings. Want last week's headlines turned into a plan for your own environment? Visit www.kinsoft.com.au to talk through your security and IT needs. Sources: Apple Newsroom; Engadget; CNBC; NPR; TechCrunch; Tech Startups; BuildFastWithAI; Europol; Google/Mandiant; CISA.

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Right now, your smartphone is getting smart enough to organize your life just by reading the context of what's on your screen. But at the exact same time... A forgotten 10 -year -old piece of human resources software just brought an entire university system to a grinding halt. Yeah, it's a staggering contradiction, honestly. We are watching this bleeding edge of futuristic technology just collide head -on with the rusted -out pipes of legacy IT infrastructure. Right. And, well, welcome

to a brand -new Deep Dive. We have a completely fascinating stack of sources to go through today, covering the week of June 8th to the 14th. It's a busy week. It really is. And the mission for today is to explore that wild whiplash between those two realities. So for you listening, whether you're a business leader trying to figure out how to safely evaluate new AI tools for your team, or you're just an insanely curious learner wanting to know what is actually happening under the hood of the devices you use every day. Which is everyone right now. Exactly. This Deep Dog is going to arm you with the essential knowledge

you need this week. And, you know, when you look closely at the facts across... all of our sources this week, there is a very clear overarching theme, one that connects the glittering AI updates to the massive cyber attacks. And that theme is supply chains. OK, let's untack this, because when most people hear supply chains, they think of cargo ships and warehouses, right? Yeah. So how does that apply to software? Well, we really need to think about digital supply chains. What we're seeing in the sources today is this deep

connection between the data supply chains that power our cutting edge AI assistance and the well, the criminal supply chains operated by modern ransomware gangs. Oh, wow. Because whether you're a tech giant trying to serve up the smartest artificial intelligence or you're a cyber criminal trying to extort millions of dollars, your success depends entirely on knowing exactly who holds your data. Right. how it flows from point A to point B, and exactly where a backdoor was left open along the way. I love that framing. So let's just jump right into the front end of that digital

supply chain, starting with the June 8th Apple Worldwide Developers Conference. Siri has finally been overhauled. We've been waiting for this integration, and it's now branded as Siri AI, running via Apple Intelligence. The demos they showed off were a massive leap forward from the basic voice commands we're used to. Yeah. The shift here is moving from a discrete tool that just like sets timers to an intelligence layer that is deeply embedded into the operating system

itself. It understands the context. Exactly. It understands the context of your entire digital life. Yeah. I mean, you can tell Siri to find a specific photo of a person from years ago and drop it into a shared family album all by voice. You can ask it to. remind you the exact minute a ticket lottery opens up. You can even take actions based on what's currently displayed on your screen, which implies it's constantly analyzing your visual workflow. Plus, they rolled out a new standalone Siri app that actually remembers the context of your past conversations. And they're

baking this intelligence directly into Spotlight Search on the Mac. Here is where the architecture of this system gets genuinely surprising. The underlying model infrastructure. Yes. Apple confirmed that this new highly capable Siri is actually powered in part by Google's Gemini model. Yeah. It's happening through Apple's Cloud Foundation model setup. And I just I have to push back on this strategy because it. It feels so completely jarring for this specific company. It did seem

a bit out of character at first glance. Right. I mean, Apple built its entire modern brand identity on the concept of total privacy and vertical integration. They designed their own silicon. They write their own software. Leaning on their biggest rival's AI for their flagship assistant. Yeah. It's like finding out the secret ingredient in a Michelin star vegan restaurant's signature dish is actually sourced from a fast food burger joint. That is a phenomenal analogy. And the irony is on. unavoidable for sure. But if we look at the mechanics of how I actually functions.

this decision makes perfect sense how so well apple is basically utilizing a peered supply chain for user queries so when you ask siri to do something highly personal like finding a photo or summarizing an email that processing happens locally on the device because that's where your sensitive data lives right exactly or it goes to apple's proprietary private cloud keeping the personal data locked down but when you ask siri a broad complex question about world history or, you know, ask it to generate a complicated

piece of code. It is more juice. Yeah. Apple's internal models just might not have the sheer breadth of general knowledge required to answer accurately without hallucinating. Building a model that knows everything about everything requires scraping the entire Internet. Which Google has already done. Exactly. Google has already done that with Gemini. So Apple basically acts as a secure router. They securely hand off that specific non -personal query to Google's infrastructure. So Apple is essentially admitting that right now, Acting as a broker for the best AI is safer than trying to build the entire encyclopedia

themselves. Pretty much, yeah. What does this mean for the listener who's trying to implement AI in their own workflow? It establishes a crucial baseline. If a company with the virtually unlimited resources of Apple cannot build a completely self -contained, perfect AI ecosystem in -house, businesses shouldn't expect to either. You're always going to be relying on third parties. Right. So the question you must ask when evaluating any AI tool is, where is this model actually running? You might purchase an enterprise license from company X, but your proprietary business

data might be getting routed and processed on company Y. Oh, that's a huge blind spot. It is. You have to map out that data supply chain before you deploy the tool. And the reason Apple has to hand off those massive general knowledge queries to Google is because of the staggering computing power required to run them, which seamlessly connects to the next massive shift in our sources. The physical hardware supply chain. We are watching a multi -billion dollar land grab for AI chips. The physical iron underneath all of this digital

magic, right? Because you can't build a software supply chain without the hardware to support it. Right. Alphabet Google's parent company. has officially started renting out its eighth -generation custom AI chips, known as TPUs, and they are operating at an unbelievable scale. Massive. They formed a $5 billion joint venture with Blackstone specifically to lease server capacity to outside companies like Anthropic and Meta. We should probably clarify what a TPU actually is just because it explains why this

hardware race is so fierce right now. Yeah, go for it. So a TPU is a tensor processing unit. For years, the tech industry relied on GPUs graphics processing units to train AI. Because GPUs were really good at doing many simple math problems at the same time. Right, which was originally needed for rendering video game graphics. But a TPU is custom designed from the ground up by Google specifically for the complex neural network math that drives machine learning. It strips away all the legacy graphics architecture. to focus purely on AI efficiency. Wow. And Google isn't the only giant building custom silicon

here. Microsoft just rolled out its second generation Maya accelerator, the Maya 200, which is specifically built to run their co -pilot and open AI workloads. Now, the sources do point out that the vast majority of Azure AI still runs on NVIDIA hardware. NVIDIA spent this exact same week tightening its grip on the hardware supply chain across Asia. They're still the king right now. Right. But the trend is clear. Hyperscalers like Google and Microsoft are desperately trying to build their own chips to escape NVIDIA's absolute dominance of the

market. Yeah, the margins on NVIDIA hardware are astronomical and the hyperscalers want to control their own destiny. But this is where I get confused by the corporate strategy here. If Microsoft and Google are spending billions of dollars in R &D to finally build their own custom chips and gain an edge. Yeah. Why is Google immediately turning around and renting those eighth generation TPUs out to direct competitors like Meta and Anthropic? I mean, if you finally have the best engine, why rent it to the other race cars? Why not hoard that advantage? It fundamentally comes down to the economics of hyperscale cloud

computing. Okay. The infrastructure costs associated with building these data centers, you know, the physical land, the custom silicon, the massive cooling systems, the energy grids, they're so astronomically high that no single company can absorb them purely for internal use. Not even Google. Not even Google. Plus, these chips depreciate quickly as... newer generations are invented. You cannot afford to let them sit idle for even a second. So by renting them out, Google turns a massive capital expense into a highly lucrative

revenue stream. So they realize the real money isn't just in building the best AI. It's in being the landlord for everyone else's AI. Exactly. They control the foundation of the physical supply chain. And for you, the end user, this dynamic is incredibly positive. How so? This fierce competition, Google, Microsoft, and Nvidia all battling to offer the most efficient compute power will ultimately drive down the cost of processing. A diversified hardware supply chain means better pricing and faster innovation for the software we actually

use. And that rapidly expanding pool of rental hardware is exactly what is supercharging the companies trying to dethrone the current AI kings. Yeah. The cheaper and more available compute power becomes, the faster a challenger can iterate, which explains the explosive growth of a major contender highlighted in our sources, Anthropic's Claude. Anthropic has been uniquely positioned to take advantage of this hardware availability to scale their operations. The numbers in the report are just staggering. According to the

data, Claude's worldwide web visit share was sitting around 8%. But then... They experienced a 300 % jump in a single quarter. Huge. Yeah, they surged from roughly 200 million visits in January to over 800 million in April. In just three months, they quadrupled their traffic. That kind of scaling isn't accidental. It points to a deliberate architectural and business shift. Anthropic realized that to capture the real value in this market, they couldn't just build a clever consumer chatbot. Right. They had to build highly efficient models with massive context windows

that businesses could actually use to process hundreds of pages of internal documents at once. And they are aggressively capitalizing on that enterprise focus. The sources note that Anthropic just formalized a $100 million partner program. They're calling it the Partner Hub. And it includes a dedicated services track specifically aimed at helping businesses put Claude into production. This really solidifies for me that the AI -assisted market is no longer just about chat GPT. No, not at all. It feels like we have transitioned

overnight from the concept car phase of AI where everyone is just marveling at the technology straight into mass -producing sedans for enterprise fleets. The novelty phase is officially over. And the key phrase in the sources that we need to analyze is putting it into production. OK, unpack that. Vendors like Anthropic are no longer just handing a business a web login for their employees to use. They are actively pushing to integrate their AI via APIs directly into a company's internal workflows. So they're essentially just

embedding themselves right into. Into the daily operations. Yeah, exactly. Their customer service staff boards and their proprietary databases. And this serves as a critical warning for anyone listening. If your team is trialing AI right now, the marketplace of available tools is widening rapidly, and the pressure to integrate them will be intense. Oh, for sure. But when you put AI into production, you are actively stitching a new powerful entity into your internal data supply chain. Conversations about data handling, retention

policies, and security access cannot be an afterthought. They must happen on day one before a single API key is generated. Security from day one. is such a crucial mindset. Because while the entire tech world is utterly mesmerized by integrating these cutting -edge screen -reading AI tools, the back door of the enterprise is being left wide open. Absolutely. We are building these highly complex, futuristic supply chains, but ignoring the rusted locks on the doors we already have. Which brings us to the dark, unglamorous reality of this week's

news. The cyber -criminal element thrives on that exact distraction. There's an extortion crew known as Shiny Hunters. Google's threat intelligence team Mandiant actually tracks their activity under the designation UNC 6240. And from late May right up to June 9th, this crew exploited an unpatched flaw in Oracle PeopleSoft. For context, Oracle PeopleSoft is a classic foundational enterprise application used for human resources, finance, and student administration. Very old school. It is the definition of legacy infrastructure.

And shiny hunters used a known unpatched flaw in this software to break into enterprise systems, steal vast amounts of data and demand extortion payments. The sources specifically note that universities were hit the hardest by this campaign. Right. On top of that, on June 10th, CISA, the U .S. Cybersecurity and Infrastructure Security Agency, added three more actively attacked vulnerabilities to its known exploited list. And, you know, when a vulnerability lands on the CISA known exploited list, it isn't theoretical. It means criminals. are actively weaponizing it in the wild right

now to breach networks. But the contrast here is making my head spin. We literally just spent the first half of this deep dive talking about billions of dollars spent on custom silicon and AI models that can visually analyze a user's desktop. Yeah. And yet massive multimillion dollar institutions like universities are being completely taken down simply because someone didn't click update on an enterprise HR app. How does a sophisticated institution fall to something so basic? It is

baffling, but it highlights the incredibly unglamorous nature of cyber defense. To understand why universities are hit so hard, you have to look at how they operate. Very decentralized. Incredibly decentralized IT departments, massive legacy user bases of students and alumni, and hundreds of web -facing portals. It makes tracking down every single instance of an old application a nightmare. Criminals do not need hyper -advanced AI to hack you. They don't need custom Google. TPUs. They just use the exact same boring old playbook they have relied on for over a decade. Walk us through

that playbook. How is it actually happening? It's a highly mechanized four -step process. Step one, use automated scanners to scour the internet for an exposed internet -facing enterprise app. Step two, the moment a vulnerability is publicly announced, race to exploit it before the target's IT team has time to test and apply the patch. Just a race against the clock. Exactly. Step three, move laterally through the network to exfiltrate the most sensitive data possible. And step four, lock the systems and extort the

victim. It is entirely opportunistic. They are simply looking for the weakest link in your software supply chain. Which means the technological defense against this massive threat is literally just patching. Yes. But if the defense is that conceptually simple. And these highly organized gangs are this relentless. How do we actually stop them? Law enforcement agencies cannot possibly run around playing whack -a -mole, forcing thousands of individual universities to update their HR software. No, they cannot patch the Internet by force, which is why global law enforcement

has fundamentally shifted their strategy. They are no longer just going after the software vulnerabilities. They're attacking the business model itself. They're targeting the financial supply chain. Following the money. Our sources detail a massive operation on June 12th by Europol. They executed a major disruption of a cryptocurrency laundering service that was used specifically by ransomware gangs to wash their illicit profits. Huge win. And the scale of this operation is hard to comprehend. This single service had claimed an estimated

€336 million. That is an industrialized scale of money laundering. €336 million of extorted money washed clean and integrated back into the legitimate economy. How does the service even begin to hide that much digital currency? Well, it requires a sophisticated mechanism to break the blockchain's chain of custody. Cryptocurrencies like Bitcoin feature public ledgers, meaning anyone can track a ransom payment from the victim's wallet to the hacker's wallet. Right. It's all public. Exactly. So to actually spend that money,

criminals use laundering services that employ mixers or tumblers. How do those work? Basically, they take the illicit cryptocurrency from... hundreds of different crimes and they pool it together. Like dumping it all into one massive pot. Right. And then they break it into thousands of microtransactions, rapidly swap it across different types of cryptocurrency chains, and then deposit the equivalent value back to the criminals. Oh, wow. Yeah. It obfuscates the origin so the funds can be cashed out for fiat currency

safely. And shutting down that Tumblr service is a massive blow. But of course, as one threat vector is disrupted, another immediately rises to take its place. Always. The sources also mention a relatively new extortion crew that has quickly climbed the ranks to become one of the most active ransomware gangs globally by victim count. And they call themselves the gentlemen. It is a striking departure from typical threat actor branding. It really is. I was noting the bizarre contrast in hacker naming conventions. On one hand, you have a flashy, almost comic book name like shiny

hunters. Right. And on the other, you have the weirdly polite, sophisticated sounding the gentlemen. But. Regardless of what they call themselves or what branding they adopt, they're all operating to feed a multi -hundred million euro... criminal ecosystem. Yeah. The naming conventions are just marketing. What is fascinating here is how clearly this Europol bust illustrates the core economics of ransomware. Yeah. Cutting that cryptocurrency money pipeline is the absolute most effective way to take the profit incentive out of cybercrime.

If the gangs cannot wash the money, they cannot spend the money and the entire operation collapses. Which is the ultimate goal. Exactly. For the listener, the enduring lesson here is that the cast of characters will constantly rotate from shiny hunters to the to whoever emerges next month. But the exploitation of unpatched systems and the underlying financial motivations remain exactly the same. It truly all comes back to the ply chains. So to seamlessly wrap up our journey today, we have explored how Apple's brand new context -aware Siri AI evolution is actually

relying on... Google's foundation model supply chain to handle complex queries. We dug into the multi -billion dollar hardware race where giants like Google and Microsoft are building custom silicon -like TPUs and immediately renting them out to control the physical infrastructure of AI. We saw Anthropix Quad leverage that compute power to achieve a massive 300 % surge in traffic as they push to integrate AI directly into enterprise production. And we contrasted all of that glittering future tech with the gritty, enduring reality

of unpatched Oracle software bringing down universities and multi -million euro pole takedowns designed to sever the financial supply chains of cybercrime. It is a complex web of information, but seeing how the physical hardware, the AI software, and the criminal networks all intersect gives us a much clearer picture of the landscape. Absolutely. So for you listening, your immediate actionable takeaways for this week are crystal clear. First, know exactly what legacy enterprise apps you have exposed to the open internet right now. Have someone check the cease and unexploited

list today and prioritize that patch queue. Second, when you are evaluating all these incredible new AI tools for your team, ask the hard questions about where your data actually lives and who else in the supply chain is processing it. And I want to leave you with a final provocative thought to mull over as you step away from this deep dive. Let's hear it. Throughout our discussion today, we have seen tech giants pooling immense resources to centralize unimaginable amounts of corporate data to feed their AI models and partner ecosystems. Meanwhile, highly organized

ransomware syndicates like The Gentleman are perfecting the art of extracting data from forgotten, unpatched systems. Right. If the future of our business productivity requires us to feed our most intimate internal data into a complex, constantly shifting supply chain of rented AI chips and third -party partner programs, are we inadvertently building the ultimate, perfectly centralized payload for tomorrow's extortionists?

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Iris Ohyama - A Suspicious File in the Cloud, 7.7 Million Customers at Risk, and an Online Store That Went Dark

    Nov 19, 202620 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Japanese home goods company Iris Ohyama (online store operator Iris Plaza) found a suspicious file in an external...

    Credit the people on this episode

  2. Who Gives A Crap - A Marketing Supplier Breached, Six Days to Tell the Client, and Customer Data That Didn't Need to Be There

    Nov 17, 202619 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Australian online retailer Who Gives A Crap told customers on 7 November 2026 that their information was accessed...

    Credit the people on this episode

  3. Last Week in Tech - An Outlook Zero-Day That Leaks Your Password Hash, Chrome's Eighth Exploited Bug, and Nvidia Without Jensen Huang

    Nov 15, 202621 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Last week in tech, 7 to 14 November 2026. Microsoft's November Patch Tuesday (10 November) fixes well over a...

    Credit the people on this episode

  4. Chainalysis - From "Non-Sensitive" to "All Users": How a Breach Notice Changed in Five Days

    Nov 12, 202619 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Chainalysis disclosed on 30 October 2026 that an external actor using "offensive agentic techniques" accessed...

    Credit the people on this episode

  5. NSW Health - A Dental Voucher Website, Four Gigabytes Taken, and the Providers Whose Details Were Exposed

    Nov 10, 202623 minAudio

    Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! NSW Health reported unauthorised access to its Oral Health Fee For Service Scheme website on 29 October 2026;...

    Credit the people on this episode